Reinforcement Learning for Multi-Period Resource Allocation Tasks
MCML Authors
David Winkel
Dr.
* Former Member
Abstract
David Winkel
Dr.
* Former Member
Abstract
This dissertation develops reinforcement learning methods for resource allocation problems, with a particular focus on financial applications and constrained decision-making. It introduces new approaches for handling linear allocation constraints with guaranteed feasibility and proposes an efficient investment strategy framework that generalizes across different risk preferences, outperforming existing methods on real-world financial data. (Shortened.)
BibTeXKey: Win25a